I'm a Supervision Analyst at the European Central Bank with a background in Statistical Learning and Mathematical Engineering (Politecnico di Milano).
The projects below are from my studies and show the statistical and machine-learning foundations I build on: nonparametric and Bayesian methods, statistical modelling, and time-series analysis, mostly in Python and R.
- Bachelor / Master of Science in Mathematical Engineering - Statistical Learning at Politecnico di Milano
- Master's Thesis — Spatial Vine Copulas for Environmental Hazard Detection: Applied advanced copula models to capture spatial dependence in high-dimensional satellite (SAR) data and identify anomalous patterns as early warning signals for volcanic activity in the Campi Flegrei area.
- Nonparametric analysis of US dairy production and consumption: This project presents a nonparametric analysis of the US dairy market dynamics, spanning from 1985 to 2021, utilizing a range of statistical methods including GAMs, bootstrap, conformal prediction, spatial analysis, and Bayesian clustering to uncover insights into optimal operating areas, pricing strategies, and trending products.
- Stochastic Block Model Prior with Ordering Constraints for Gaussian Graphical Models: Constructed a Gibbs Sampler from scratch and introduced a new Bayesian prior within the framework of Gaussian Graphical Models enabling the learning of conditional dependence block structure among variables, while accommodating ordering constraints.
- Development of a wearable fall detection system using Machine Learning: Created a Python-based classification system that utilizes accelerometer data to differentiate various activities of daily living, including running, walking, and falling. This system leverages machine learning models and employs time series feature extraction techniques for accurate activity recognition.
- E-mail: puri.andrea@libero.it
- ♟ I enjoy playing chess
- 🎾 Tennis player for 10+ years
- 🐲 Love watching TV series